The mask set for a shuttle run (multi-project wafer) may contain designs using different number of metal layers. Wafers
fabricated with k metal layers can only yield dice for the designs using only k metal layers. This results in considerable
waste of wafers. In this paper we propose a simple concept called Lifted-I/O (LIO) to address this problem. LIO elevates
the I/Os of all designs to the highest metal layer possibly used in a shuttle run. Our study shows that a shuttle run with
LIO for low-volume production is marginally better than that without LIO. However, the margin increases as production
volume increases. For a production volume up to a few ten thousand dice per design, a shuttle run with LIO could pay up
to 11% less, which corresponds to about 400 thousand dollars.